In search of knowledge: text mining dedicated to technical translation

Although a vast amount of contents and knowledge has been made available in electronic format and on the web in recent years, translators still do not have friendly and targeted tools at their disposal for the various aspects of a translation process, i.e., the analysis phase, automatic creation and management of the linguistic resources needed and automatic updating with the relevant information generated by the computer translation tools used in the process (Machine Translation, Translation Memories, and so on). Text mining and information retrieval are not typically connected with the translation process and no existing online translation workspace integrates text mining or information retrieval facilities that are specifically aimed at improving the documentary competence of translators in order to process unstructured (textual) information, and make the information on the web or in texts accessible to translators. This paper explores a new approach to helping translators look for different types of information (glossaries, corpora, Wikipedia, and so on) related to the specific translation work they have to perform which can then be used to update the lexical base needed for the translation workflow (both human or machine-aided). This new approach is based on CATALOGA, a text mining tool, which can be combined with an IR application and/or an MT/TM system and used for different purposes. 1 Johanna Monti is author of the Abstract, Introduction, Sections 2 and 4.3 and Conclusions, Annibale Elia is author of Sections 3.1 and 3.4, Alberto Postiglione is author of Section 3.2. and 5, Mario Monteleone is author of Sections 3.3 and 4.1 and Federica Marano is author of Sections 4.2. and 4.4.